J. exp. Biol. 139, 317-328 (1988) 3 1 7'Printed in Great Britain © The Company of Biologists Limited 1988

ELECTROPHYSIOLOGY OF NEUROSECRETORY CELLSFROM THE PITUITARY INTERMEDIATE LOBE

BY ROBERT N. McBURNEY1 AND STEVEN J. KEHL2

Cambridge NeuroScience Research, Inc., 1 Kendall Square, Cambridge, MA,USA and 2Department of Physiology, University of British Columbia,

Vancouver, BC, Canada

Summary

One of the goals in studying the electrical properties of neurosecretory cells is torelate their electrical activity to the process of secretion. A central question inthese studies concerns the role of transmembrane calcium ion flux in the initiationof the secretory event.

With regard to the secretory process in pituitary cells, several research groupshave addressed this question in vitro using mixed primary anterior pituitary cellcultures or clonal cell lines derived from pituitary tumours. Other workers,including ourselves, have used homogeneous cell cultures derived from thepituitary intermediate lobes of rats to examine the characteristics of voltage-dependent conductances, the contribution of these conductances to actionpotentials and their role in stimulus-secretion coupling.

Pars intermedia (PI) cells often fire spontaneous action potentials whosefrequency can be modified by the injection of sustained currents through therecording electrode. In quiescent cells action potentials can also be evoked by theinjection of depolarizing current stimuli. At around 20°C these action potentialshave a duration of about 5 ms. Although most of the inward current during actionpotentials is carried by sodium ions, a calcium ion component can be demon-strated under abnormal conditions. Voltage-clamp experiments have revealed thatthe membrane of these cells contains high-threshold, L-type, Ca2+ channels andlow-threshold Ca2+ channels.

Since hormone release from PI cells appears not to be dependent on actionpotential activity but does depend on external calcium ions, it is not clear what rolethese Ca2+ channels play in stimulus-secretion coupling in cells of the pituitarypars intermedia. One possibility is that the low-threshold Ca2+ channels are moreimportant to the secretory process than the high-threshold channels.

Introduction

Our ideas about the process of neurotransmitter release at synapses in both theperipheral and central nervous systems are well formed and supported by a verylarge body of evidence. The arrival of a Na+-dependent action potential at apresynaptic nerve terminal causes the brief opening of voltage-dependent Ca2+

Key words: neurosecretion, pars intermedia, electrophysiology, ion channels.

318 R. N. MCBURNEY AND S. J. K E H L

channels in the surface membrane of the terminal and the resulting influx ofcalcium ions is the necessary and sufficient condition for initiating the processwhereby vesicles containing the chemical transmitter substance discharge theircontents into the synaptic cleft (Katz, 1969; Kandel, 1985). Although this scheme,which links a depolarizing event to the activation of Ca2+ channels, applies tosynapses and perhaps to some secretory cell types which generate actionpotentials, such as chromaffin cells (Douglas, 1968), it is far from clear whether itapplies to the secretory cells of the pituitary gland. Results obtained in somepituitary cell types (e.g. ovine gonadotrophs, Mason & Waring, 1985) suggest that,although hormone release stimulated by releasing factors depends on externalCa2+, Na+-dependent action potentials might not play an essential role in thisprocess.

One difficulty in investigating stimulus-secretion coupling in pituitary cells isthat the gland itself contains several different secretory cell types that secretedifferent peptide hormones in response to a variety of stimuli, generally unique toeach cell type. If one wants information about the cellular mechanisms underlyingthe secretory process in one class of pituitary cell it is necessary to develop apreparation which contains just one cell type or in which particular cell types canbe identified so that the appropriate stimuli can be applied to the cell(s) underinvestigation and, for release studies, so that an assay can be made for theappropriate hormone. The application of techniques such as immunohisto-chemistry and the reverse haemolytic plaque assay (see other chapters in thisvolume), density centrifugation (Cobbett, Ingram & Mason, 1987) and fluor-escence-activated cell sorting (Chen, St John & Barker, 1987) to the identificationof cell types is now making it possible for us to gain insight into the stimulus-secretion coupling process in particular anterior pituitary cell types. Furthermore,the revolution in the electrophysiological investigation of small cells in vitroafforded by the introduction of whole-cell and patch-clamp recording techniques(Hamill et al. 1981) allows us to gain detailed information about the electricalcharacteristics of individual secretory cells that might be relevant to stimulus-secretion coupling.

An alternative approach to obtaining a homogeneous population of a pituitarycell type is to use a clonal cell line (such as GH3 cells, Tashjian, Bancroft & Levine,1970) with concomitant problems in relating the results obtained to any non-tumour celi type of the pituitary. One further tactic that has attracted a smallnumber of groups is to use an anatomically identifiable region of the pituitarygland which contains a homogeneous population of hormone-secreting cells. Onesuch region is the pars intermedia (PI) or intermediate lobe of the rat pituitary. Itis readily dissected free of the anterior and posterior lobes and yields a populationof cells known as melanotrophs whose secretory products include a--melanocyte-stimulating hormone (a'-MSH), corticotropin-like intermediate lobe polypeptide(CLIP), and a variety of other products of the precursor molecule pro-opiomelanocortin (Eberle, 1981).

In addition to containing a homogeneous population of hormone-secreting cells

PI cell electrophysiology 319

this region of the pituitary gland is unique in that the secreting cells are innervated.Histochemical studies have shown that the terminals of this innervation contain awide variety of neurotransmitters and neuropeptides and from this finding onemight infer that the cells of the pars intermedia are subjected to a far morecomplex control regime than anterior pituitary cells. However, since we knowrelatively little about the functions of the pars intermedia and its hormones, it isdangerous to draw any such conclusions. Nevertheless, because this region of thepituitary can be isolated and studied in vitro, either intact or as dispersed cells, it isan excellent source for an identified pituitary cell type which has the bonus ofpossessing a wide range of neurotransmitter receptors and their signal transduc-tion mechanisms.

Following the lead of Douglas, Taraskevich, their colleagues and a smallnumber of other groups, we have studied the electrophysiological properties ofsingle cells from the rat pituitary pars intermedia which were maintained indispersed cell culture (for methods see Tomiko, Taraskevich & Douglas, 1984;Cota, 1986; Kehl, Hughes & McBurney, 1987). This article is a brief overview ofthe current state of knowledge about the electrophysiological properties of parsintermedia cells and is a commentary on present understanding of stimulus-secretion coupling in these cells.

Pars intermedia cells in cell cultureThe process of dispersing the intermediate lobe produces small 'rafts' of cells

and isolated cells (Fig. 1). The isolated cells are about 15(Um in diameter and verysuitable for use with whole-cell recording techniques (Cota, 1986; Kehl etal. 1987).The advantages of using isolated cells in cell culture as opposed to acutely isolatedglands for both electrophysiological studies and hormone release experiments arethe good visibility of the cells for electrophysiology and the fact that ininvestigations of electrically stimulated and/or releasing-factor (transmitter-induced cellular events one can be confident the action of these stimuli is only onthe secretory cells and not on any aspect of their innervation (see for exampleTomiko, Taraskevich & Douglas, 1984; Taraskevich, Tomiko & Douglas, 1986).The major disadvantage of this in vitro preparation is the loss of innervation,which may lead to the expression of abnormal cellular properties (Cota &Armstrong, 1987).

Fig. 1. Light micrographs taken under phase-contrast optics of cells derived from thepituitary intermediate lobe of adult rats and maintained in monolayer culture. Lobeswere broken up, as described in Kehl, Hughes & McBurney (1987), into clumps of cellsand single cells then plated onto a confluent layer of rat cerebral cortical astrocytes(phase-dark background cells). Some of the cells flattened on the background layerlosing their phase-bright appearance (A) whereas other cells, both isolated and in'rafts', maintained an approximately spherical shape (B). The cells we have used formost experiments were isolated, phase-bright cells like the one in the lower left cornerof B. Scale bar, 50jum.

320 R. N. McBURNEY AND S. J. KEHL

rmrtfrr-fTrrrnrTfifTfOmV

Fig. 2. Spontaneous activity recorded from a PI cell in the cell-attached (upper trace)and whole-cell (lower trace) configurations. Prior to applying a pulse of suction toobtain a whole-cell recording, the patch current showed biphasic waveforms reflectingspontaneous firing in the cell. Subsequently, in the whole-cell configuration, the restingpotential fluctuated between -50 and -35 mV, the latter potential being the thresholdfor the activation of action potentials which overshot the 0 mV level by 30-35 mV. Thepeak amplitude of the action potentials shown here was attenuated by the low-frequency response of the pen recorder. The basis for the shift from a tonic (cell-attached record) to a phasic pattern (whole-cell record) of firing is unknown, butperhaps is the result of a change in the intracellular environment of the cell. The solidhorizontal line in the lower trace record denotes the OmV level.

'Resting' properties of cultured pars intermedia cellsEarly experiments, using both intracellular and extracellular recording tech-

niques, demonstrated that isolated pars intermedia cells possess spontaneouselectrical activity (Douglas & Taraskevich, 1978). This activity consists of sub-threshold fluctuations in the cell's resting membrane potential and actionpotentials (Fig. 2). Because the input resistance of the cells is so high (around4GQ, S. J. Kehl & R. N. McBurney, unpublished observations) it is possible thatthe voltage fluctuations arise as a result of the opening and closing of single ionchannels in an asynchronous manner. If this is so both the cause of this channelactivity and the nature of the ion channels involved remain unresolved.

Action potentials in pars intermedia cells

The characteristics of action potentials in these cells are relatively straightfor-ward. They are brief depolarizing events which can overshoot OmV by 30-35 mVand are followed by a marked after-hyperpolarization (Fig. 3). At room tempera-ture they have a half-width of about 5 ms and under these conditions a monotonicrepolarization phase, unlike some rat sensory neurones which reveal the calciumcomponent of their action potential as a hump in the repolarization phase (Dichter& Fischbach, 1977). In some cells which show spontaneous voltage fluctuationsno action potential activity, the injection of sustained depolarizing current can

PI cell electrophysiology 321

4pA f-|_|4pA

100 ms

OmV-

,10pA

Fig. 3. Injection of depolarizing (A-E) and hyperpolarizing currents (F) into aquiescent PI cell. Each panel shows a current injection trace above a membranevoltage trace. The dashed line represents OmV for the membrane voltage trace. Theresting potential for this cell was — 60 mV. The amount of injected current is shown atthe end of each current trace. The biphasic deflections on the current traces are due toinadequate control of the membrane current during action potentials. Note that thefrequency of action potentials is related to the magnitude of the current injection. Notealso the marked accommodation of the action potentials in C, D and particularly in E.

generate action potentials whose average frequency depends on the magnitude ofthe injected current (Fig. 3).

Relationship between action potential activity and secretion

When the intermediate lobe is isolated from its innervation the averagefrequency of action potentials in the secretory cells rises and hormone outputincreases. From this observation it has seemed reasonable to propose that actionpotentials are the important event in the secretory process and that the regulationof hormone output from the gland in vivo is by the modulation of a tonic inhibitionmediated by the dopamine- or y-aminobutyric acid (GABA)-containing fibres ofits innervation. If this proposal were true one might expect that the basal secretion

prom isolated glands or cells would be suppressed in the presence of tetrodotoxin(TTX). Unfortunately TTX does not suppress basal secretion from either isolated

322 R. N. McBURNEY AND S. J. KEHL

150

2 ioo

3O

X 502

I

0TTX Ca2+-free

Fig. 4. Basal melanocyte-stimulating hormone (MSH) output from isolated melano-trophs or neurointermediate lobes is not inhibited by tetrodotoxin (TTX) but isinhibited by omission of Ca2+. Average responses from preparations of dispersed ratpars intermedia cells (N=4; empty columns), mouse pars intermedia cells (N = 3;stippled columns) or mouse neurointermediate lobes (7V= 2; hatched columns); errorbars denotes S.E.M. or the range. The values plotted are the outputs of MSH during thelast 5min of a 10-20 min exposure to TTX or Ca2+-free solutions. These values areexpressed as a percentage of the basal secretion during the 5 min that preceded theintroduction of TTX or the omission of Ca2+. The responses illustrated in Ca2+-freesolution were obtained subsequently from the same preparations on which TTX hadbeen tested. Reproduced from Tomiko, Taraskevich & Douglas (1984) with per-mission.

cells or isolated neurointermediate lobes (Fig. 4). However, hormone release isreduced in Ca2+-free bathing solutions (Fig. 4).

This pair of observations was the first sign that the release process in parsintermedia cells is not completely straightforward. One interpretation is that theaction potentials are not associated with any membrane conductance change inresponse to Ca2+ and that the calcium entry which affords the release process adependence on extracellular [Ca2+] is through Ca2+ channels which are notprimarily associated with action potentials. Although this idea is attractive and hasa parallel in ovine gonadotrophs (Mason & Waring, 1985), there is no directevidence to support this for rat pars intermedia cells. Furthermore, the actions ofneurotransmitters such as dopamine, GABA, serotonin (5-HT) and noradrenalineon hormone output are associated with changes in the spontaneous actionpotential activity of pars intermedia cells (Douglas & Taraskevich, 1982;Taraskevich & Douglas, 1982, 1984; Tomiko, Taraskevich & Douglas, 1983) and itis possible to demonstrate the existence of voltage-dependent Ca2+ channels inthese cells which might contribute a component to normal action potentials(Douglas & Taraskevich, 1980; Cota, 1986; Hughes, Kehl & McBurney, 1987|Kehl, 1987).

PI cell electrophysiology 323

OmV lOmV

|lOOpA

10 ms|lOOpA

10 mV 10 mV

jlOOpA

Fig. 5. Voltage-activated ionic currents in rat PI cells obtained with the whole-cellrecording technique. All recordings were obtained by stepping the membrane voltagefrom -70 mV to the potential indicated at the end of each current trace then steppingback to — 70mV. Without going into too much detail, the following solutionmodifications were made to achieve the recordings shown. (A) 'Normal' membranecurrent response to a voltage step. No membrane ion channels intentionally blocked.(B) Inward current isolated by substituting Cs+ for K+ in the recording electrodesolution. (C) Na+ current revealed by the addition of 300,umol 1~' Cd2+ to the externalsolution while blocking outward currents with internal Cs+ as in B. (D) Ca2+ currentrevealed by the addition of 2//moll"1 tetrodotoxin (TTX) to the external solutionwhile blocking outward currents as in B. (E) Fast transient K+ current revealed byblocking Na+ and Ca2+ currents with TTX and Cd2+, respectively, and additionallyblocking the delayed-rectifier K+ current (see F) with 20mmoir' tetraethyl-ammonium in the external solution (S. J. Kehl, unpublished observation). The internalsolution contained K+ as the predominant cation. (F) The delayed-rectifier K+ currentrevealed by blocking both Na+ and Ca2+ currents by the methods indicated above.This cell did not display a fast transient K+ current.

The voltage-dependent ion channels in rat pars intermedia cells

Fig. 5 illustrates the variety of membrane ionic currents that we have been ableto record from isolated cells under various experimental conditions. This summaryfigure shows that conventional Na+ channels and delayed-rectifier K+ channels arepresent (Fig. 5C,F, respectively) and that inward Ca2+ currents can be elicited by adepolarizing voltage step (Fig. 5D). Fig. 5 illustrates the presence of a fasttransient K+ conductance mechanism in some PI cells (S. J. Kehl & R. N.|WcBurney, unpublished observations). Our findings largely confirm and extendthe work of Taleb, Trouslard, Demeneix & Feltz (1986) on porcine PI cells and the

324 R. N. MCBURNEY AND S. J. KEHL

V(raV)-30 mV

50 mV

40 mV

30 mV

20 mV

25pA

10 ms

I(pA)

Fig. 6. The properties of Ca2+ currents in rat PI cells. (A) Ca2+ currents recorded bystepping the membrane potential from a holding potential of — 70 mV to variousmembrane potentials indicated with the current traces. Leakage currents have notbeen subtracted from these records. (B) I/V relationship for Ca2+ currents measuredat the end of 25 ms voltage pulses in a rat PI cell (from Cota, 1986). Note thecharacteristic shape of the I/V curve which is very similar to that shown by Carbone &Lux (1984) and indicates the contributions of two separate Ca2+ channel types to therecorded currents. Reproduced from Cota (1986) with permission. (C) I/V relation-ship for peak Ca2+ currents we have recorded from rat PI cells. The maximum currentlevel of the I/V plot is similar to that shown in A but a clear separation of the twocomponents which have been proposed to contribute to the currents is not seen.

work of Cota (1986) on rat PI cells. Clearly voltage-dependent Ca2+ channels arepresent and could conceivably contribute to action potentials and stimulus-secretion coupling in these cells.

Properties of the voltage-activated Ca current

Cota (1986) has suggested that the voltage-activated Ca2+ current in pars

PI cell electrophysiology 325

intermedia cells is generated by the activation of two separate ion-channelcomponents defined in terms of their thresholds for activation and the speeds atwhich they deactivate after the voltage pulse has returned to the holding potential.These two components have been named the low-threshold, slowly deactivating(SD) channel and the high-threshold, fast-deactivating (FD) channel (Fig. 6B).When the properties of these Ca2+ channels are compared with those of Ca2+

channels described in vertebrate sensory neurones (Carbone & Lux, 1984) it isclear that, in terms of activation thresholds, the two classes of Ca2+ channel aresimilar in the two different cell types. However, there may be some differences inthe inactivation properties of the channels between the two preparations. Moreexperimental work will certainly be needed before proper comparisons can bemade amongst different cell types on this feature of the voltage-dependent Ca2+

channels.The current-voltage curves (Fig. 6C) which we have derived from measuring

peak calcium currents during voltage-clamp steps (Fig. 6A) do not show such aclear separation of low-threshold and high-threshold conductances. Perhaps ourmethodology has prevented us from clearly identifying this component. Neverthe-less, our results confirm the observations of a sizeable voltage-dependent Ca2+

conductance in PI cell membranes. Both investigations show that this conductancebegins to activate around — 50 mV and is maximally activated around OmV.

One feature of the Ca2+ conductance change that is worth noting is therelatively slow activation kinetics following a step change in membrane voltage(Fig. 6). This slow activation means that during a brief (5-10 ms) action potentialnot all the available Ca2+ channels will be activated. In fact, depending on theprecise relationship between the kinetics of the various ionic conductances thatcontribute to the action potential, the fraction of the total number of Ca2+

channels that is activated might be very small, especially in the case of the high-threshold channels. Nevertheless, the inward Ca2+ current resulting from theactivation of a very small fraction of available Ca2+ channels might raise thecytosolic concentration of free calcium ions to levels sufficient to trigger hormonerelease directly or to evoke additional Ca2+ release from an intracellular store.

The exact relationship between action potentials, the activation of Ca2+

conductances and the intracellular concentration of calcium ions ([Ca2+]j) will onlybe determined by the direct measurement of [Ca2+]j during action potentials in PIcells in a manner similar to that recently reported for GH3 cells (Schlegel et al.1987).

Role of Ca2+ channel types in the secretory process

It seems unlikely that high-threshold Ca2+ channels would make any contri-bution to the secretory process if action potentials were eliminated. Once the Na+

channels have been blocked by TTX it is extremely difficult to evoke a Ca2+ spike'by the artificial injection of depolarizing current under normal conditions(Douglas & Tarakevich, 1980). However, it is possible that some of the low-

326 R. N. MCBURNEY AND S. J. K E H L

threshold Ca2+ channels are continuously opening and closing at the 'resting'potential in PI cells. Such activity might be, in part, responsible for thespontaneous voltage fluctuations mentioned earlier (see Fig. 2).

The inward Ca2+ currents underlying the fluctuating membrane voltage mightmaintain [Ca2+]s at levels sufficient to cause hormone secretion. Any transientincreases in [Ca2+]j associated with action potentials might, in overall terms, berelatively unimportant to secretion from PI cells. The scenario mentioned abovewould explain the lack of TTX-sensitivity of the release process and its depen-dence on extracellular calcium ions. However, since the preceding sentences inthis section contain one 'possible' and three 'mights' it is clear that muchexperimental work will be necessary to put this suggestion on a firmer foundation.

If the low-threshold Ca2+ channels are more involved in the secretory processthan the high-threshold Ca2+ channels then it is not clear why the stimulantdihydropyridine BAY K 8644 can markedly potentiate hormone secretion from PIcells in a TTX-insensitive manner (Taraskevich & Douglas, 1986). Since thedihydropyridines are relatively specific for the L-type, or high-threshold, Ca2+

channel in neurones (Nowycky, Fox & Tsien, 1985) one can only conclude that theCa2+ channels of PI cells are not identical in their pharmacological properties tothose previously studied in neurones.

The possible greater importance of low-threshold Ca2+ channels to the secretoryprocess than high-threshold channels has implications for the design of exper-iments to investigate the modulation of secretion from PI cells. If it is necessary touse K+ depolarizations to evoke additional secretion from cells, the K+ concen-trations used should be chosen in the light of the activation curves for the variousCa2+ channel types.

ConclusionsElectrophysiological studies are revealing the spectrum of voltage-dependent

ion channels that are present on the surface membranes of secretory cells of thepituitary intermediate lobe. The role of these ion channels in stimulus-secretioncoupling in PI cells is not clear. Future experiments, where electrical events,[Ca2+]j and secretion are measured from single cells (see, for example, Penner &Neher, 1988), will undoubtedly provide much of the information needed to resolvesome of the paradoxes which abound in the literature. Whether the finalmechanism of stimulus-secretion coupling in PI cells is unique or can begeneralized to other pituitary cell types remains to be determined.

Our work on PI cells was performed at the MRC Neuroendocrinology Unit,Newcastle upon Tyne, England. SJK held a postdoctoral training fellowship fromthe Canadian MRC. We would like to thank D. Hughes, J. Slade, H. Brown andW. K. Gascoigne for valuable technical assistance.

PI cell electrophysiology 327

ReferencesCARBONE, E. & Lux, H. D. (1984). A low voltage-activated, fully activating Ca channel in

vertebrate sensory neurones. Nature, Lond. 310, 501-502.CHEN, G.-G., ST JOHN, P. A. & BARKER, J. L. (1987). Rat lactotrophs isolated by fluorescence-

activated cell sorting are electrically excitable. Molec. cell. Endocr. 51, 201-210.COBBETT, P., INGRAM, C. D. & MASON, W. T. (1987). Sodium and potassium currents involved in

action potential propagation in normal bovine lactotrophs. J. Physiol., Lond. 392, 273-299.COTA, G. (1986). Calcium channel currents in pars intermedia cells of the rat pituitary gland.

Kinetic properties and washout during intracellular dialysis. J. gen. Physiol. 88, 83-105.COTA, G. & ARMSTRONG, C. M. (1987). Long term stimulation of dopamine receptors inhibits

calcium channel activity in cultured pituitary cells. Soc. Neurosci. Abstr. 13, 794.DICHTER, M. A. & FISCHBACH, G. D. (1977). The action potential of chick dorsal root ganglion

neurones maintained in cell culture. J. Physiol., Lond. 267, 281-298.DOUGLAS, W. W. (1968). Stimulus-secretion coupling: the concept and clues from chromaffin

and other cells. Br. J. Pharmac. 34, 451-474.DOUGLAS, W. W. & TARASKEVICH, P. S. (1978). Action potentials in gland cells of rat pituitary

pars intermedia: inhibition by dopamine, an inhibitor of MSH secretion. J. Physiol., Lond.285, 171-184.

DOUGLAS, W. W. & TARASKEVICH, P. S. (1980). Calcium component to action potential in ratpars intermedia cells. J. Physiol., Lond. 309, 623-630.

DOUGLAS, W. W. & TARASKEVICH, P. S. (1982). Slowing effects of dopamine and calcium-channel blockers on frequency of sodium spikes in rat pars intermedia cells. J. Physiol., Lond.326, 201-211.

EBERLE, A. (1981). Structure and chemistry of the peptide hormones of the intermediate lobe.In Peptides of the Pars Intermedia (ed. D. Evered & G. Lawrenson), pp. 180-190. London:Pittman Press.

HAMILL, O. P., MARTY, A., NEHER, E., SAKMANN, B. & SIGWORTH, F. J. (1981). Improved patch-clamp techniques for high-resolution current recording from cells and cell-free membranepatches. Pflugers Arch, ges. Physiol. 391, 85-100.

HUGHES, D., KEHL, S. J. & MCBURNEY, R. N. (1987). Patch-clamp analysis of the intrinsicmembrane properties of adult rat melanotrophs maintained in vitro. J. Physiol., Lond. 391,80P.

KANDEL, E. R. (1985). Factors controlling transmitter release. In Principles of Neural Science(ed. E. R. Kandel & J. H. Schwartz), pp. 120-131. New York: Elsevier.

KATZ, B. (1969). The Release of Neural Transmitter Substances. Springfield: Thomas.KEHL, S. J. (1987). A high-threshold non-inactivating calcium current in isolated rat

melanotrophs in vitro. J. Physiol., Lond. 391, 74P.KEHL, S. J., HUGHES, D. & MCBURNEY, R. N. (1987). A patch clamp study of y-aminobutyric

acid (GABA)-induced macroscopic currents in rat melanotrophs in cell culture. Br. J.Pharmac. 92, 573-585.

MASON, W. T. & WARING, D. W. (1985). Electrophysiological recordings from gonadotrophs.Evidence for Ca2+ channels mediated by gonadotrophin-releasing hormone.Neuroendocrinology 41, 258-268.

NOWYCKY, M. C , Fox, A. P. & TSIEN, R. W. (1985). Three types of neuronal calcium channelwith different calcium agonist sensitivity. Nature, Lond. 316, 440-443.

PENNER, R. & NEHER, E. (1988). The role of calcium in stimulus-secretion coupling in excitableand non-excitable cells. J. exp. Biol. 139, 329-345.

SCHLEGEL, W., WlNlGER, B. P., MOLLARD, P., VACHER, P., WuARIN, F., ZAHND, G. R.,WOLLHEIM, C. B. & DUFY, B. (1987). Oscillations of cytosolic Ca2+ in pituitary cells due toaction potentials. Nature, Lond. 329, 719-721.

TALEB, O., TROUSLARD, J., DEMENEIX, B. A. & FELTZ, P. (1986). Characterization of calciumand sodium currents in porcine pars intermedia cells. Neurosci. Letts 66, 55-60.

TARASKEVICH, P. S. & DOUGLAS, W. W. (1982). GABA directly affects electrophysiologicalm properties of pituitary pars intermedia cells. Nature, Lond. 299, 733-734.TARASKEVICH, P. S. & DOUGLAS, W. W. (1984). Electrical activity in adenohypophyseal cells and

effects of hypophyseotropic substances. Fedn Proc. Fedn Am. Socs exp. Biol. 43, 2373-2378.

328 R. N. MCBURNEY AND S. J. KEHL

TARASKEVICH, P. S. & DOUGLAS, W. W. (1986). Effects of BAYK8644 and otherdihydropyridines on basal and potassium evoked output of MSH from mouse melanotrophs invitro. Neuroendocrinology 44, 384-389.

TARASKEVICH, P. S., TOMIKO, S. S. & DOUGLAS, W. W. (1986). Electrical stimulation ofneurointermediate lobes of mice elicits calcium dependent output of melanocyte-stimulatinghormone. Brain Res. 379, 390-393.

TASHJIAN, A. H., BANCROFT, F. C. & LEVINE, L. (1970). Production of both prolactin and growthhormone by clonal strains of rat pituitary tumor cells. J. Cell Biol. 47, 61-70.

TOMIKO, S. A., TARASKEVICH, P. S. & DOUGLAS, W. W. (1983). GABA acts directly on cells ofpituitary pars intermedia to alter hormone output. Nature, Lond. 301, 706-707.

TOMIKO, S. A., TARASKEVICH, P. S. & DOUGLAS, W. W. (1984). Effects of veratridine,tetrodotoxin and other drugs that alter electrical behaviour on secretion of melanocyte-stimulating hormone from melanotrophs of the pituitary pars intermedia. Neuroscience 12,1223-1228.

VINCENT, J. D., DUFY, B., ISRAEL, J.-M. & ZYZEK, E. (1984). Electrophysiological approach ofstimulus-secretion coupling in prolactin-secreting cells. In Prolactin Secretion (ed. F. Menna& C. Valverde-R.), pp. 209-224. New York: Academic Press.